Self-report validity issues.

Two possible sources of the substantial gap usually found between survey self-reported alcohol consumption estimates for a population and estimates based on official alcohol sales records are investigated. A measure of atypical heavy drinking is added to ordinary consumption commonly measured in surveys, and consumption by an adolescent (age 14-17) sample is added to that of the adult sample. The relationship between respondents' purchases and consumption during a 30-day period is also investigated. Personal interviews were completed with a random sample of 997 adults and 182 adolescents in Iowa during February-April 1985. Adding atypical drinking to ordinary drinking narrowed the sales-self-report gap more than did adding adolescent drinking, but a considerable gap remained. Self-reported purchases were closer to sales than was self-reported consumption. However, not all purchasers were drinkers and not all drinkers were purchasers, and the two were not highly correlated. The self-report validity issue, which remains unresolved, is apparently affected by many factors. Self-reports appear to be accurate enough for some purposes but not for others. Official alcohol sales (or purchase) records are not necessarily valid measures of alcohol consumption.